Nanofactory Products A to Z

What could a nanofactory make? [PDF]

That's the question that J. Storrs (Josh) Hall answered during his talk at CRN's recent nano/bio conference in Tucson. This week we're highlighting a few of the conference presentations that have been uploaded here.

Josh began by offering an interesting comparison between the falling costs and corresponding increase in power of computing over the past 40 years, and the projected costs and power estimates for fabrication devices over a similar period. (Click images below to enlarge.)

Then Josh described various categories of items that might be produced by a personal nanofactory. The first three were Apples, Bicycles, and Cups of Coffee -- see any pattern here?

Among the more interesting potential nanofactory products that Josh talked about were Money, Robots, and Utility Fog. Each of those items suggests all sorts of fascinating implications.

Tomorrow: "Seeing the nano future through storytelling."

Mike Treder

Tags: nanotechnology nanotech nano science technology ethics weblog blog

Comments

[Nato Welch]

I was surprised to see him him talk about food so much; I had the impression that foodstuffs would be among the last, most complex things MM could produce.

What about drugs? I did a quick survey of various drugs (prescription and illegal), and it seems they, also, conform to being mostly CHON. In fact, all of the Illegal drugs I looked at were only CHON, with the prescription drugs usually needing S or F.

Never mind bombs - is the "war on drugs" going to serve as sufficient justification to keep nanofactories out of public hands?

[Svein Ove Aas]

This conference looks more interesting every day that passes. For those of us who really had no possible way to be there, won't there be audio recordings at some point?

[Sigma]

Interesting stuff, although I have to agree with Nato Welch that he focuses way too much on food. Also his highlighting of mundane objects was kind of odd, perhaps he was trying to get through to the layman... who knows.

I was wondering if any of you guys had heard about recent developments in microfactories? While not as capable as a true nanofactory, they could definently be disruptive, especially if it can be self-replicating.

http://www.eurekamagazine.co.uk/article/9914/Manufacturing-base-keeps-on-shrinking.aspx

[Mike Treder, CRN]

I probably should have specified that Josh was talking not just about a first-generation nanofactory -- which presumably will have a capacity too limited to produce foods -- but about the range of products that could be made by future nanofactories. He was giving us a broad sweep.

As for audio of the talks, we're working on it, and hope to have something up within the next week or so.

[Josh Hall]

Like any other technology, nanofactories will be subject to user-pull as well as technology-push. We see a lot more vending machines today for food than for gadgets. The current Fab@Home can make edible objects. It doesn't synthesize the food from CHON but it doesn't synthesize plastic either.

Even though food is more complex than simple gadgets, I think the market will keep food synthesis more or less even with the other kinds in practice.

In the talk, I pointed out that the major nutrients -- protein, carbohytrates, and fats -- can be provided in a reasonably limited set of basic feedstock molecules (e.g. the 20 amino acids) which can be manipulated by a somewhat easier chemistry than general molecular synthesis. For example, your feedstock could be a solution of a specially designed protein. There would be a special-purpose disassembler for exactly that protein, that broke it into aminos that could in turn be re-sequenced into a wide variey of proteins for various tastes and consistencies. (but that you're still nowhere near building life-like cell structures).

Outputs not limited to food (think silk, keratin).

Like a personal computer, a personal nanofactory has an almost unimaginably wide range of possibilities; but to take off in the market there must be a few "killer apps". Food and clothing are obvious starts. Any other ideas, anyone?

--Josh

ps -- one "killer app" isn't a product but the stage where the dissasembler gets good enough to handle the household waste stream.

[Dan S]

CHON is obviously insufficient for making food. Leaving aside S, there are minor (but important!) amino acids like selenocystein (contains Se), and DNA/RNA synthesis needs P. What about metal ions such as Na Mg,Mn Ca...?

>>For example, your feedstock could be a solution of a specially designed protein. There would be a special-purpose disassembler for exactly that protein, that broke it into aminos that could in turn be re-sequenced into a wide variey of proteins for various tastes and consistencies.<<

No need for "specially designed protein"/"special-purpose disassembler" here - this stage could employ natural proteases for breaking down any protein. These are already contained in gastric juice. As for "various tastes" - food taste is determined by small molecules not by large proteins and their amino acid sequences and various aromatizers already used in food industry.

Therefore in my opinion described device is not necessary and not sufficient for producing food. It can't enhance neither nutritional qualities of food (since all amino acids must be present in feedstock) nor taste (since taste is not a function of amino acid sequence).

>one "killer app" isn't a product but the >stage where the dissasembler gets good >enough to handle the household waste stream

Thats looks like general-purpose high throughput dissasembler which is very dangerous device – one of main components of “gray goo”. I remember that CRN position about such dissasemblers devices was always “it was never proposed by MM proponents” and “completely unnecessary”.

[Josh]

Actually taste can be determined by amino acid sequence, and across a wide range of molecular sizes. For example, the proteins thaumatin, monellin, mabinlin, pentadin, brazzein, curculin and miraculin have a sweet taste, and range from 6 to 98 kDa. As for disassemblers, you are conflating a microscopic self-replicating autonomous machine (bad) with what's basically an improvement to your garbage disposer, such that it has an extra "out" pipe, providing feedstock back to the synthesizer (good).

[Chris Phoenix, CRN]

It's not clear to me that a molecular-level disassembler will ever be an efficient way to deal with random macro-scale trash streams. I could see sorting rotors (molecular binding and transport mechanisms) being used to purify and/or concentrate mixed material streams.

But sorting rotors wouldn't work well for very large random molecules. (This doesn't rule out protein-sorting medical applications, because there are fewer proteins and you can spend more energy on purifying them.) And to analyze a very large molecule and figure out how to pull it apart would probably take more energy for computation than would be used in simply burning it and sorting the small gas molecules with sorting rotors.

For construction, give me molecular manufacturing (almost) every time. For deconstruction of unknown materials, give me bulk chemistry.

Chris

[Dan S]

Josh,
>>Actually taste can be determined by amino >acid sequence, and across a wide range of molecular sizes.<<

Ok, I think I can see you point here. Some proteins apparently can bind taste receptors thus producing effect stronger than small molecules. Still, I have a number of questions:
1. As far as I understand these proteins are exceptions. Normally food taste is determined more by small molecules and raw chemical composition than by precise protein structure. Thus raw amido-acid compositions together with other small molecules can get good approximation to natural tastes. What are new abilities added by resequencer, that are not already available?
2. Do you propose wide usage of taste-receptor interacting proteins in food? Are there enough natural proteins to cover all taste ranges?
3. To make new functional protein right amino-acids sequence is not enough. Many proteins undergo post-translational modification (miraculin for example contains disulfide bonds). How such cases could be handled by resequencer?
3. Furthermore to achieve their function, these proteins must fold correctly. While some proteins fold spontaneously, correct folding with high yield is only achieved inside living cells. It depends on local environment of ribosome during translation and sometimes other proteins are needed. What about health-safety of such device? Eating misfolded proteins can sometimes be a cause of disease.

>>you are conflating a microscopic self-replicating autonomous machine (bad) >with what's basically an improvement to your garbage disposer<<

I fully agree that garbage disposer augmented with disassembler technology is not dangerous by itself. But technology needed to build such disposer and autonomous microscopic replicator is basically the same, isn’t it?

Chris,
Bulk chemistry is not always superior to molecular disassembly. Not when you want process to go at room temperature and pressure and/or not produce toxic pollutants. Living systems do use disassembling (proteases, nucleases...). Why MM, being superior to living systems in performance could not be used to make effective disassemblers? Sort molecules into several classes by their composition. Then apply predefined deconstruction procedure to all molecules of particular class. Are you absolutely sure that systems that work in such way are impossible?